This study will test the hypothesis that an increased randomness of ACTH and cortisol secretion is present in patients with head injury. Head injury includes patients who have sustained closed-head injuries from trauma such as motor vehicle accidents and vascular accidents such as strokes and hemorrhages. A common complaint of head injury patients is an increase in fatigue and lethargy. This hypothesis is based on initial studies in patients admitted to the Transitional Learning Community in Galveston, TX. Out of 44 patients studied, 50% (22) had low morning serum cortisol concentrations although they did respond normally to ACTH stimulation. Disruption in neural inputs to the hypothalamus and pituitary from head injury may alter the normal secretin patterns of ACTH and correspondingly cortisol. This subtle alteration cannot be detected by standard stimulation testing of the axis, but we will determine in this study whether an algorithm that measures approximate entropy (ApEn) can determine this randomness. This mathematical model detects differences in underlying hormone episodic behavior using vector scaling to assess the amount of randomness between consecutive hormone samples in a time series. Differences in ApEn will be determined between control subjects, head injury patients with normal baseline serum cortisol concentrations, and head injury patients with below normal cortisol concentrations. Blood samples for ACTH and cortisol will be drawn every 15 min overnight (8pm to 8am) with continuous monitoring of sleep stages. In this 2 year study, sixteen subjects will be studied in each group with an equal number of men and women and an age range of 20-35 years. Control subjects will be age and gender-matched. If ApEn can show randomness of hormone secretion in head injury patients, then treatment paradigms to correct this abnormal secretion can be developed and the improvement in functional status of the patients assessed.